Abstract This study achieved high-performance depth-of-interaction (DOI) encoding in PET detectors through systematic optimization of crystal surface reflection processing, utilizing a light-sharing redirection mechanism. We employed a 23 × 23 LYSO array coupled to an 8 × 8 SiPM array under a 9:1 crystal-to-photodetector coupling ratio. Three detector designs with different surface treatments were investigated: Detector I (1.02 × 1.02 × 15 mm 3 crystals, fully polished with E60 film), Detector II (1 × 1 × 15 mm 3 crystals, ESR-wrapped with unpolished sidewalls), and Detector III (1 × 1 × 15 mm 3 crystals, fully polished with ESR film). Detector I achieved the best DOI resolution (3.29 mm FWHM), while Detector II reached 4.13 mm, both detectors can achieve good crystal separation after layering, enabling effective discrete identification. Notably, the Detector III design enabled direct crystal discrimination without layered structures but exhibited compromised depth sensing capability. Energy resolution in the central region was 16.22%, 13.72%, and 11.54% for Detectors I–III, respectively, and the corresponding timing resolutions were 1.45 ns, 1.02 ns, and 1.64 ns. Taking all aspects into consideration, Detector II exhibited the most balanced performance, indicating that selective surface texturing (e.g., sidewall unpolishing) combined with high-reflectivity materials (ESR) can effectively optimize photon transport pathways. This methodology enables efficient DOI information capture while maintaining compatibility with existing PET electronics, providing a technically viable solution for next-generation high-resolution PET systems that integrates performance advantages with engineering feasibility.
Yuan et al. (Wed,) studied this question.